Thanks very much for the reply! Did you create a complete database (with alignments, trees etc) or just the fasta sequence and mapping file? It would be great if you could give a general overview of steps to achieve this!
I have a special case where I want to use my custom 16s data (already classified to respective species) to be used with QIIME. Does anyone know how can I create a greengenes like database for my data (to be used as Closed-reference OTU picking pipeline)?
Thanks very much for any help you can offer!
1 answer
Because Greengenes is rather limited with Archaea, I recently made a QIIME compatible version of SILVA 119 nr99.
You just need a taxonomy mapping file ID<TAB>kingdom; phylum; class; order; family;genus (note 6 levels if you use RDP classifier, for other assignment methods maybe all the way down to species level).
ID1 k__Archaea; p__Euryarchaeota; c__; o__; f__; g__
ID2 k__Archaea; p__Euryarchaeota; c__; o__; f__; g__
ID3 k__Archaea; p__Euryarchaeota; c__; o__; f__; g__
..
And then your sequences with corresponding headers, in this case ID1, ID2, ID3. For some QIIME analyses you would also need alignment of your sequences and a tree. I made my reference with a few awk one-liners and GNU coreutils (basically just parsed the taxonomy info from the sequence headers and made sure that it was down to genus level in all cases with just "g__;" representing missing genus info and so forth).
I didn't need alignments or trees so I didn't bother. With SILVA, I could have just grabbed the alignment from their site though. I believe with Greengenes they used ssu-align for the alignment and FastTree for tree building. All the info is in the README files..
Thank you for pointing out the readme file. I see they have brief methods section in there. I'll try to replicate that!
Log in to answer this question.